A web form that drops each new lead into the CRM and emails the team is a simple workflow, and published rate surveys put that at $500 to $2,000.
A custom AI agent that finds, researches, scores and drafts runs $5,000 to $15,000. Keeping either one running usually costs a few hundred to a few thousand dollars a month. What moves the number is scope, not the consultant's title.
The figures below come from two kinds of source: sellers' own pricing pages, and published rate surveys of what operators and agencies charge (Vantaige and betonai). Each range carries the name of whoever published it. Prices were checked in October 2026.
Price ranges by type of engagement
| What you're buying | Published range | Source |
|---|---|---|
| Productized small-business AI audit | $250 to $1,000 | Vantaige 2026 rate survey |
| Mid-market AI audit | $1,500 to $5,000 | Vantaige |
| Consultant-led readiness audit (2 to 4 weeks) | $3,500 to $8,500 | Main & Machine, published pricing |
| Single n8n workflow, one-time | $400 to $1,200 | BULDRR AI, published pricing |
| Simple workflow | $500 to $2,000 | Vantaige |
| Integrated workflow across several tools | $2,000 to $5,000 | Vantaige |
| Several workflows, one-time | $1,500 to $4,500 | BULDRR AI |
| Custom AI agent or multi-workflow system | $5,000 to $15,000 | Vantaige |
| Implementation sprint (4 to 12 weeks) | $18,000 to $60,000 | Main & Machine |
| Enterprise build | $15,000 to $75,000+ | Vantaige |
What AI automation agencies charge per month
| Ongoing work | Published range | Source |
|---|---|---|
| Monitoring for one or two automations | $300 to $1,500 a month | Vantaige |
| Optimization, actively improving the system | $1,500 to $5,000 a month | Vantaige |
| n8n “care and feeding” | $650 to $1,200 a month | betonai rate survey, 54 operators |
| Managed services | from $1,500 a month | Main & Machine |
| Automation retainer | $1,200 to $8,000 a month | BULDRR AI |
One line on BULDRR's page is worth reading twice: the client pays the n8n subscription and the API bills on top. That's normal across the field, and it's the first thing to ask about.
AI consultant hourly rates and fractional leadership
| Who | Published rate | Source |
|---|---|---|
| Individual automation operators | $75 to $200 an hour | Vantaige |
| Senior operators and specialists | $150 to $300 an hour | Vantaige |
| Architecture-level and fractional CTO work | $300 to $500 an hour | Vantaige |
| One solo AI consultant's published rate | $500 an hour | Ryan Doser |
| Fractional AI leadership, small business | $3,500 to $8,000 a month | Ryan Doser, citing Tavoko |
| Fractional AI leadership, mid-market | $8,000 to $30,000 a month | Ryan Doser, citing Tavoko |
If the job is answering the phone
AI phone answering gets its own table, because its pricing works differently from everything above.
| Option | Published price | Source |
|---|---|---|
| Self-serve AI receptionist software | from $29 a month | Dialzara |
| Self-serve AI receptionist software | $49 to $299 a month, by minutes | Rosie |
| Self-serve AI receptionist software | $79 to $299 a month, by unique callers | Goodcall |
| Self-serve AI receptionist software | $99 a month for 200 minutes, then $0.25 a minute | My AI Front Desk |
| Self-serve AI receptionist software | free for 25 calls, then $3 a call; Pro $150 a month at $2 a call | Smith.ai |
| Done-for-you starter voice agent | $1,500 to $2,500 setup, plus $300 to $450 a month | one agency's published guide (Ciela) |
| Human answering service | $175 to $700 a month | OnceHub |
| A front-desk hire | $2,800 to $4,500 a month, fully loaded | beancount.io cost guide |
Watch the unit. Some plans meter minutes, some count callers, some charge per call. A plan that looks cheap at 200 minutes can look very different in a busy month.
What actually drives the price
How many tools have to talk to each other. One form into one CRM is a small job. A lead that has to move through the website, the CRM, the calendar and the accounting software is a bigger one, and every hand-off is a place it can break.
How clean the data is. If customer records live in three places with three spellings of the same name, somebody has to sort that out before anything gets automated on top of it.
Whether it talks to customers. Anything a customer sees carries higher stakes than an internal report. It needs a designed way out for the customer, which is more work to build and test.
Who runs it after launch.Somebody watches it, fixes it when a tool changes underneath it, and adds the next piece. That's the monthly line, and it's real work.
Whether the process works by hand yet. Automating a bad process just gets you bad results faster. Automating one you've already proven works gets you good results faster. If the process isn't settled, the money goes into settling it first.
The costs that don't show up on the quote
Overages. Per-minute and per-call pricing with no cap is how a monthly bill doubles in a good month.
Platform and API fees.Automation platforms and AI models bill separately from whoever builds on them. Ask whose card they're on.
Staff cleanup.A cheaper tool gets expensive fast if someone on the team still has to re-enter customer information or patch a broken hand-off every week. That time never appears on an invoice, which is exactly why it's easy to miss.
How to tell if it's worth it
Two formulas cover most cases.
For time: hours a week × what that person costs per hour with taxes and benefits × 4.3 = the monthly cost of doing it by hand.
Example numbers, not a client: six hours a week at $35 an hour is about $900 a month.
For missed work: leads a month that never get a response × the share that would have booked × what a job is worth.
Example numbers, not a client: four unanswered leads, half of them likely to book, at $600 a job, is $1,200 a month.
Put either number next to the monthly cost and the setup fee spread over a year. If the automation doesn't clear it comfortably, it's the wrong piece to automate first.
Questions to ask before you pay anyone
- Will you write down what you'd automate, and what it should save, before any build?
- Did they name a tool before they understood the problem?
- Is the price fixed, and how do changes get agreed?
- What does it cost to run after launch, and whose account do the software and API bills go on?
- If an AI talks to my customers, what can a customer do when it can't help them?
- Do I approve each step before the next one starts?
How I price
Every engagement starts with a conversation. Then about a week to diagnose: I watch how the work moves and write down what's worth automating and what isn't. Each step after that gets a written, fixed price before it starts, and any change is agreed in writing first.
The first paid step is a small, fixed-price prototype of the one piece worth proving, running on your own process, over about two weeks. A Louisville recruiting firm did exactly that: the prototype pulled open job listings, researched each company, scored it against the kind of client they wanted and drafted the outreach for their team. If a prototype proves out, the build takes two to six weeks. After that I either run it month to month or train the person on your team who'll own it. You approve each step before the next one starts.
More on each step: how working with me works. If you're in Louisville or Southern Indiana: AI automation in Louisville.